Falsely elevated troponin I results due to outliers indicate a lack of analytical robustness

Abstract

Aims Troponin (Tn) is the preferred biochemical marker for the diagnosis of acute coronary syndrome. Spurious false Tn results (outliers) may cause significant problems with clinical management. We investigated the occurrence of outliers and whether this phenomenon could be explained by analytical imprecision. Methods and results Troponin I (TnI) was measured in duplicate with Beckman AccuTnI reagent if the first TnI result was ≥0.04 μg/L ( n = 5265). All TnI requests were performed in duplicate in a subset of samples for one calendar month ( n = 881). A total of 13,178 TnI requests were received during the study period. Variables were sample type, centrifugation speed and analyser. Results were identified as outliers when the difference between two results exceeded a critical difference (CD) limit defined by CD = z × √2 × SDAnalytical. Outliers at the 0.0005 probability level were detected in 102 of 5265 duplicate observations (1.94 ± 0.37%). This translated into an outlier rate of 0.55 ± 0.13% for all TnI results and 1.37 ± 0.31% for results above 0.04 μg/L. Outliers resulted only in falsely elevated TnI values and were not dependent on the analyser, centrifugation speed or sample type. Conclusion TnI outliers occurred more frequently than anticipated, could not be explained by analytical imprecision and indicated a lack of robustness in the assay. The high rate and the magnitude of the errors will complicate clinical management and carry a risk of detrimental patient outcome. The outlier rate is a useful parameter to define the robustness of assays.